[illumos-Developer] Review for 719 beadm should allow BEs outside of <rpool>/ROOT

Alexander Eremin alexander.eremin at nexenta.com
Wed Mar 30 00:18:12 PDT 2011


On Mar 30, 2011, at 5:59 AM, Richard Lowe wrote:

> be_create.c:627
> be_snapshot.c:298
>
>  This block is, I think, duplicated, could it be factored out?

Correct, thanks

>
> be_utils.c:540
>
>  I think splashimage.xpm is correct, but I may have been the person
> to cause this confusion.
>
>  I commented on the prior round that use of 'splash.xpm.gz' seems
> wrong, because that's almost certainly the wrong splash xpm (I have no
> idea why it's delivered, or what the confused history there is).
> 'splashimage.xpm' should be correct, and 'solaris.xpm' should be a
> link to it, to maintain compatibility.  Using 'splashimage.xpm' would
> be correct.  References to 'splash.xpm.gz' would, I think, not be.

But we haven't /boot/splashimage.xpm in newly created  BE - only
solaris.xpm and grub fails in my tests

>
> libbe_priv.h:45
>
>  Perhaps it's just me, but "RPOOL_CONTAINER" seems like a crummy
> name, not at all indicative that it's actually a boolean
>
> general:
>
>  You should update the beadm manual page to describe these changes.
>
>  Bill Sommerfeld asked in the first review why you needed two
> properties, rather than having the setting of BENAME_STARTS_WITH imply
> the value of RPOOL_CONTAINER

Correct

>
>  I asked whether this was really generally useful, or something which
> would be better kept in Nexenta's own tree, as it seems almost
> entirely useful to maintain compatibility with their existing
> filesystem layout.
>
> -- Rich

I ran into this before Nexena. Why not add beadm to this opportunity.  
As I said earlier,
maybe not all the distributions in the future will use the IPS and  
its ​​parameters.
I would like to have a choice.

Thanks,
:: Alexander









More information about the Developer mailing list